8:47
A little over a month ago, just as the South Korean stock market was collapsing due to the hype surrounding AI, the government made a big statement that got a little lost in the noise of why 10x leverage on tech stocks is a bad idea. The government has announced that it will provide free access to AI for all its citizens without restrictions on use, and the first wave of such universal computing power is set to be tested by the end of this week. Of course, in an economy as reliant on AI as South Korea's, there's an argument that this is yet another form of circular finance for which the industry is being criticized, only on a national scale. Besides, AI is quite expensive, and basic social economics teaches us that people don't
value things they don't have to pay for. So with unlimited usage in a country with a population of 51 million that continues to shrink, this could quickly turn into a very expensive bill. But is this really a bad idea? Given other challenges facing Korea's economy, such as an aging population and shrinking workforce, technology that is ultimately designed to replace human labor could prove beneficial to an economy with a shortage of workers. And this step has a certain historical basis. Korea is not the first country to publicly bet on a new trendy technology. Some of these early steps have provided countries with decades of productivity growth, while others have been a waste of resources and, to some extent, a disgrace. And while no
one, especially economists, can predict the future, it's worth asking a few questions to understand where this experiment might lead, because, one way or another, it seems likely that similar programs will be implemented in our economies, for better or for worse. So how would free AI for everyone in South Korea actually work? What will prevent this project from becoming a huge burden on taxpayers? And finally, can it really fix the underlying problems of their economy? Free AI in South Korea will have many consequences, but one we can't ignore is a surge in entrepreneurship. According to today's issue sponsor, Shopify, entrepreneurship has added over $850 million to South Korea's GDP in the last 4 years alone
. AI is not the only reason for this growth, but it is safe to say that these numbers will only grow thanks to the free use of AI. This is not just happening in South Korea, it is happening all over the world. Since 2022, economic activity on the Shopify platform has more than doubled globally, reaching $1.2 trillion, adding $687 billion in economic activity in the US alone this year. And with over 5.6 million new business registrations last year, this trend is only set to continue. At the heart of it all is Shopify, a universal commerce platform that lets you get your idea to market faster than ever.
Similar to South Korea, Shopify has its own AI assistant called Sidekick, available to all users. It helps with almost every task and further shortens the path from idea generation to market launch. While all of these features are great, the real advantage of Shopify is that they root for you. They want you to build the business you've dreamed of and help you every step of the way. Despite the general negativity about the future, it's nice to work with a brand that is genuinely optimistic about it. If you want to check out the new Shopify for Entrepreneurs report, the link will be in the description or go to shopify.com/entrepreneur.
The results of the study are truly inspiring. So, " free AI for all with no token restrictions" is a big headline, but despite the "universal basic computing" label that some try to attach to it, the real program is much more modest than it initially seemed. The program consists of two main parts. The first is an AI system designed to help people interact with government services. This includes, for example, obtaining a passport, applying for social programs, or completing bureaucratic procedures that usually require the assistance of The idea is that instead of waiting 40 minutes to ask a tax return question, you simply talk to an AI that can pull up all the records and give you an answer on the spot.
This system will be connected directly to government databases and specifically trained to solve problems that support staff in these positions typically deal with. Honestly, this part is unlikely to be radically different from the chatbots already running on Amazon or almost any other site that has decided to save on real support workers. The only thing that makes this noteworthy is that the program is being implemented by the government, and even though the initiative is larger than other examples, it is not entirely unique. This is a very similar program to the one recently launched in the US, and despite the attempts of some individuals to discredit it to the administration, it is essentially just an expanded FAQ page. The second part of this
program, however, is a bit more interesting. It offers free general- purpose AI chat through select apps to everyone in the country, and the service has been rolled out by three of Korea's largest technology and telecommunications companies. Kakao, which runs the country's main super app; KT, one of the country's largest telecommunications providers; and SK Telecom, which is part of the same corporate family as SK Hynix, one of the world's largest memory chip manufacturers. The idea is that users will be able to connect with these chatbots through a number of different channels, such as the KakaoTalk app, which functions similarly to WhatsApp or WeChat, which is a fairly standard solution. But they will also be accessible
via regular phone calls and text messages, which is clearly a conscious design choice aimed at making it easier to use for the country's large and growing elderly population. As for how it's actually paid for, the government is providing these companies with a significant GPU server infrastructure to run it all-reportedly hundreds of advanced Nvidia server nodes -in addition to small annual payments for the duration of the trials. These payments are unlikely to cover the real costs of running these programs on this scale, but the program has an important condition: the majority of traffic, about 80%, must pass through Korean AI models.
This means that this is effectively an opportunity for these companies to use government subsidies to take market share away from paid American or Chinese alternatives, while the tools themselves will be limited to individual accounts tied to Korean citizens. It's also worth noting that the chatbots on offer will be relatively basic, nothing like the advanced models that can cost hundreds of dollars per month for complex queries. They will be more like an unlimited version of the free tier features that most of these AI companies already offer. But the hope is that once people are in this ecosystem and have built relationships with these models-whether through long-term memory features that remember previous conversations or, let's be honest, even
just emotionally-it will become much harder for users to abandon them. This means that they may later start paying for premium features or pay for continued use altogether when the government program ends. So, if all goes well, this is a great deal for these megacorporations, even if it won't necessarily be profitable. But why exactly is the government doing this? Now I wanted to tell you about the Economics Explained newsletter. If you're watching this video and think there should be more stories like this, we've created a newsletter for that. Once a week, we send out one economic piece about things that quietly shape the world but are rarely properly explained. The newsletter is completely free, you can subscribe via the link in the description
or via the QR code on the screen now. Let's get back to the reasons why the South Korean government is providing free access to AI. Well, government subsidizing access to new technologies isn't always the worst idea. Many countries, including Korea itself, have invested heavily in rolling out high-speed broadband internet to all citizens because they saw the potential of this technology to increase productivity across the economy. The logic was that if the technology was made accessible, businesses and individuals would find productive uses for it, leading to higher productivity, greater tax revenues, which would eventually pay for the initial costs and more. And in the
case of Korea, that's exactly what happened Businesses saw that they could connect to the world wide web and began to build services and products based on it. This created a foundation of technical expertise that helped the country become a world leader in a wide range of advanced technologies and League of Legends, which, let's be honest, is probably a statistically significant drag on productivity. However, the digital infrastructure that emerged from those early investments is a big reason why the country ranks much higher in the technology sphere than one might expect. It was like a small seed crystal around which the rest of the
new economy could develop. But if we were to develop this analogy, it only works when there is a favorable environment for crystal growth and when that crystal itself is worth growing. The Internet obviously paid off handsomely in the end, but not every government bet on new technology worked so well. Just a few years ago, we were talking about El Salvador becoming the first crypto- native economy when the government launched a national Bitcoin wallet and encouraged its adoption among the entire population. The results were, at best, mixed. The value of government assets fluctuated in exactly the way reserve assets should not, and the level of implementation among the population did not even come close to what was written in the
headlines. Of course, no one can predict the future, and it is entirely possible that this experiment will turn out to be a waste of money, but AI has already shown much greater promise in real productivity gains, especially in the intellectual labor and manufacturing sectors that are key to Korea's economy. It may not be as transformative as the internet, but all else being equal, it will likely bring much more real economic value than speculative, zero- sum assets like cryptocurrency. So, it would be fair to say that the historical precedent is not bad, and everything is planned to build domestic potential, and not just hand out money to foreign technology firms. But can this really solve the deep structural
problems of the Korean economy? In many ways, South Korea is where many of our economies will be in about 20 years if current demographic trends continue. That's why many economists are watching this country to see what happens next. We've talked about this in other videos, so I don't want to repeat myself, but the fertility rate in South Korea fell to 0.72 in 2023, and preliminary data for 2024 suggests a further decline to about 0.68, which is about a third of the level needed for population stability. The country's workforce is shrinking and the population is aging, which means that in the coming decades there will be fewer and fewer people of working age creating the economic output needed to support an increasing number of retirees. So the argument for
free AI is relatively clear. If you don't have enough employees, give those you have the tools to make them more productive. In theory, you can maintain the same level of production with fewer people. Anything that can empower workers should help reduce the impact of workforce reductions, and this logic is entirely sound, but often a bit short-sighted. For starters, South Korea already has difficulty finding stable, quality jobs for its youth. The current rate of youth disengagement, which includes discouraged workers and people stuck in involuntary part-time employment, is around 18%. A significant portion of young Koreans have simply dropped out of the labor market altogether.
Many spend years preparing for highly competitive entrance exams or taking odd jobs. It's hard to argue that South Korea is so desperate for young workers that it urgently needs AI to replace them when those same young workers are forced to go through a veritable " Hunger Games" to get an entry-level position. Ultimately, automating the lowest rung of this career ladder certainly doesn't help. Moreover, if these young people cannot find stable jobs, they are even less likely to start families and have children, which only exacerbates the very demographic problem that this technology is supposed to alleviate. From a
reputational point of view, the timing also doesn't look the best. This program was announced after the Korean stock market had already plummeted due to a combination of overly risky AI speculation, political uncertainty, and broader global trade tensions, but it was clearly planned and developed at the height of the AI investment mania, of which the government was particularly proud. Ordinary Korean retail investors were investing massively in AI and semiconductor stocks, and instead of curbing this wild speculation and protecting people from its inevitable consequences, the government prepared such statements, knowing that it would only fuel a dangerous hype.
This is provided that they did not make a complete mistake with the choice of time. Now, none of this means that the program is guaranteed to fail. There are strong arguments for making AI tools widely available, especially in an economy facing demographic challenges like Korea. But, at the very least, it would be naive to think that this was done solely with the best economic intentions. Even if we consider this narrow policy in isolation, many will be watching how this experiment unfolds. Other technology companies in other countries have already put forward the idea of universal basic computing power. So, regardless of the results of the Korean version, this concept is unlikely to disappear anytime soon. The question for the rest of us is
whether this will be a Korean broadband breakthrough, or turn into something like El Salvador's Bitcoin wallets. Although, perhaps, this is not the main thing. The integration of public services into inevitable technological platforms, the blurring of the line between public infrastructure and private profit, and the use of subsidies to tie populations to specific corporate ecosystems-all these are central concepts of technofeudalism, an economic theory that has recently gained much popularity for rather obvious reasons. We already made a whole video on this topic, so we didn't want to repeat ourselves too much here. But if you're interested, you can jump to it by clicking on the link on the screen now. Thanks for watching, friend.
Goodbye.
